Identifying colorectal metastases in liver biopsies: the novel CDX2 antibody is less specific than the cytokeratin 20+/7- phenotype.
BACKGROUND: Immunohistochemistry plays an important role in tracing the primary site in metastatic tumors of unknown origin. Therefore, determining the cytokeratin (CK) 20/CK7 pattern of metastases is one of the most helpful procedures as the CK20+/CK7- pattern is typical of colorectal adenocarcinomas. Expression of CDX2 protein is a new, highly specific and sensitive marker of the intestinal origin of adenocarcinomas. In the present study we compared the sensitivity and specificity of CDX2 expression and the CK20+/CK7- phenotype in predicting the colorectal origin of liver metastases. MATERIAL/METHODS: The study was carried out on a consecutive series of 125 core-needle biopsies of metastatic adenocarcinomas of the liver. Most of the patients were followed up to death, and primary tumor localization could be established in 102 cases by a combination of clinical, radiological, histological and, in some cases, autopsy data. All the needle biopsies were immunohistochemically stained for CK7, CK20 and CDX2. CDX2 expression (at 10% and 50% cut-off levels) and the CK20/CK7 pattern of the metastases were correlated to the primary site established. RESULTS: The CK20+/CK7- pattern showed a specificity of 98.7% in predicting colorectal primary localization, which was superior to that of CDX2 expression at both cut-off levels (90% and 95.3% respectively). The sensitivity of CDX2 expression in these circumstances was 84% at the 10% cut-off, somewhat higher than that of the CK20+/CK7- phenotype (79.5%), but lower at the 50% cut-off level (72.7%). CONCLUSIONS: The CK20+/CK7- immunophenotype is more specific in predicting the colorectal origin of liver metastasis than CDX2 expression.[1]References
- Identifying colorectal metastases in liver biopsies: the novel CDX2 antibody is less specific than the cytokeratin 20+/7- phenotype. Tot, T. Med. Sci. Monit. (2004) [Pubmed]
Annotations and hyperlinks in this abstract are from individual authors of WikiGenes or automatically generated by the WikiGenes Data Mining Engine. The abstract is from MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine.About WikiGenesOpen Access LicencePrivacy PolicyTerms of Useapsburg